Cross-modal Vision-Language Retrieval on Amstertime
13.8R@1La-SigLIP-V2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| La-SigLIP-V2Augmentation=Language-augmented (La-)2026.02 | 13.8 | 29 | 40 | 49.5 | |
| La-BLIPAugmentation=Language-augmented (La-)2026.02 | 12.8 | 31.4 | 41.1 | 53.2 | |
| La-CLIPAugmentation=Language-augmented (La-)2026.02 | 11.5 | 27.5 | 37.7 | 50.4 | |
| La-EVA-V2Augmentation=Language-augmented (La-)2026.02 | 11.5 | 27.9 | 36.8 | 47.8 | |
| EVA-CLIP-V2Augmentation=None2026.02 | 2.8 | 8 | 12.7 | 18.1 | |
| CLIPAugmentation=None2026.02 | 2 | 7.5 | 12.1 | 18 | |
| BLIPAugmentation=None2026.02 | 1.3 | 4.1 | 7 | 12.3 | |
| SigLIP-V2Augmentation=None2026.02 | 0.9 | 3.1 | 5.5 | 8.8 |